575

J Clin Pathol 1992;45:575-578

Prospective study of necropsy audit of geriatric inpatient deaths D A Paterson, M I Dorovitch, D L Farquhar, H M Cameron, C T Currie, R G Smith, W J MacLennan

Abstract Aims:To evaluate the accuracy of clinical diagnosis by specialists in geriatric medicine and to compare this with a previous study involving non-specialists. Method: Clinical and necropsy diagnoses from consecutive hospital inpatient deaths from the University Department of Geriatric Medicine were analysed for discrepancies at regular audit meetings. Three main categories of diagnosis were considered and any therapeutic implications discussed. Results: Between 1987 and 1989 necropsies were performed on 100 patients (38 men, 62 women, aged 63 to 99 years) from a total of 207 deaths, a necropsy rate of 50%/.. There was complete agreement between necropsy and clinical diagnoses in 32% of cases. Disagreement involved the main diagnosis in 28%, contributory conditions in 32%, and cause of death in 34%. In 10% of cases the diagnostic discrepancy was considered therapeutically important. Specialist geriatricians correctly diagnosed the main diagnosis in 72% of cases; non-specialists in the previous study were correct in only 47% of

Department of Pathology, University of Edinburgh Medical School, Teviot Place, Edinburgh EH8 9AG D A Paterson M I Dorovitch D L Farquhar H M Cameron C T Currie R G Smith W J MacLennan Correspondence to: Dr D A Paterson

Accepted for publication 28 November 1991

That study, however, was based on necropsies from acute medical and surgical wards served by the University Department of Pathology and did not include patients dying under the care of specialists in geriatric medicine.

Methods Data were collected on a two part proforma. In part A the clinical diagnoses were recorded before necropsy and confirmed by the consultant geriatrician responsible for the patient. Diagnostic confidence was graded as certain, fairly certain, or uncertain. Four categories of clinical diagnoses were recorded: 1 The main diagnosis-that is, the most important condition underlying the patient's last illness and death (similar to the last diagnosis listed in Part I of the current death certificate). 2 The contributory conditions, both related and unrelated to the main diagnosis (Parts Ia, Ib, and II). 3 The final cause of death (Part Ia). 4 Other diagnoses unrelated to the final illness or death. Part B was completed by the pathologist performing the necropsy, and comprised a cases. simple list of necropsy findings. Conclusion: Specialist geriatricians diagAt a subsequent clinico-pathological meetnose elderly people more accurately than ing geriatricians and pathologists discussed non-specialists. But rates of misdiagnosis each case and made a joint decision as to are still significant and necropsies con- whether or not the necropsy diagnoses could tinue to be a useful form of audit. be reconciled with the clinical diagnoses as outlined in Part A. Clinical diagnoses unlikely to result in diagnostic morphological features Numerous studies have detailed discrepancies at necropsy were accepted where they were between necropsy findings and clinical diag- supported by appropriate investigations. noses, with levels of misdiagnosis ranging from Where a necropsy diagnosis was listed clini10% to 39% for the main clinical diagnosis.` 10 cally but assigned to a different diagnostic Such discrepancies have been shown despite a category, then the necropsy allocation was high level of confidence in the clinical diag- taken as correct, but these were not recorded nosis;' and one study found that 15% of such as clinically important misdiagnoses. Where missed diagnoses were of clinical importance misdiagnosis was identified a decision was as they would have led to substantially different made as to whether this would have led to different investigations or treatments. The feapatient management.1" Achieving an accurate clinical diagnosis is sibility of achieving the correct diagnosis in life especially difficult in elderly patients, many of was discussed and recorded. whom will have multiple pathology and in The subjects reviewed came from 38 acute whom presentation of disease may be atypical. assessment and 93 long stay beds at the Compliance with clinical assessment and sub- University Department of Geriatric Medicine sequent investigations may be hindered by Unit the City Hospital, Edinburgh. In the confusion and frailty. A previous large nec- study period from November 1987 to March ropsy study in Edinburgh established that the 1989 there were 207 deaths and necropsies accuracy of clinical diagnosis declined with were requested in all cases. Permission was increasing patient age, so that only 47% of granted for 104 cases, an overall necropsy rate clinical main diagnoses were confirmed at of 50%. From these, 100 satisfactory pronecropsy in 295 patients aged more than 75.12 formas were completed. The age and sex

Paterson, Dorovitch, Farquhar, Cameron, Currie, Smith, MacLennan

576

Table 1 Distribution of 100 cases by sex and age

distribution of the cases is presented in table 1.

Age (years)

Sex

55-64

65-74

75-84

85-94

95+

Total

Male Female Total

1 0 1

2 4 6

19 28 47

15 27 42

1 3 4

38 62 100

CODING

For ease of analysis and data presentation both clinical and pathological diagnoses were coded

using the Ninth Revision of the International Classification of Diseases (ICD). Diseases of the

Table 2 Agreement of clinical diagnoses with equivalent necropsy diagnoses Diagnostic category

Agreement

Disagreement

Total

Main diagnoses Contributory condition Cause of death Others All diagnoses

72 68 66 77 32

28 32 34 23 68

100 100 100 100 100

Table 3 Over- and underdiagnosis of main diagnosis by ICD codes with diagnoses ICD Code

Overdiagnoses: I II VI

VIla

VIII Total

Category

Infective Neoplasia Nervous system Cardiovascular

Number 4 2 1 9

Respiratory

4 20

II VIIa

Neoplasia Cardiovascular

3 10

VIIb VIII IX

Cerebrovascular Respiratory Digestive

2 2 2

X XIII Total

Genitourinary Musculoskeletal

Underdiagnosis:

21

circulatory system were divided into VIIa (cardiovascular) and VITb (cerebrovascular). As the "other" diagnoses unrelated to death had no clinically relevant effect on the clinical management of the patients a detailed analysis of these is not presented. Results DISTRIaUTION OF DIAGNOSES

Cardiovascular diseases, including congestive cardiac failure, ischaemic heart disease, acute myocardial infarction and pulmonary embolus, were both the most common main diagnoses and contributory conditions. Respiratory disorders, in particular bronchopneumonia, were the most common cause of death.

Clinical main diagnosis

Septicaemia (4) Lung (1), carcinomatosis (1) Cervical compression (1) Pulmonary embolus (4), congestive cardiac failure (1), acute myocardial infarction (2), deep venous thrombosis (1), chronic myocardial ischaemia (1) Bronchopneumonia (4) Necropsy main diagnosis Bile duct (1), liver (1), colon (1) Acute myocardial infarction (3), aortic stenosis (2), pulmonary embolus (2), left ventricular failure (2), arteritis (1) Cerebral infarction (2) Aspiration pneumonia (2) Gastric ulcer (1), diverticular disease(l) Acute pyelonephritis (1) Septic arthritis (1)

COMPARISON OF CLINICAL AND NECROPSY DIAGNOSES

The rate of agreement of clinical diagnoses with necropsy diagnoses is shown in table 2. There was complete agreement on all categories of diagnosis in 32 cases. In 10 of the 68 cases of disagreement the discrepancy was regarded as of no clinical relevance as it involved only disorders unrelated to the main diagnosis. The discrepancies were of greater importance in the remaining 58 cases and were of two sorts-overdiagnosis, in which clinical diagnoses were not confirmed at necropsy; and underdiagnosis, in which there was no clinical suspicion of the necropsy diagnosis (tables 3-5). CLINICAL RELEVANCE OF THE DISCREPANCIES

Table 4 Over- and underdiagnosis of contributory conditions by ICD code with common diagnoses Number Clinical contributory condition ICD Code Category Overdiagnosis: II Neoplasia

VHa VIIb VIII IX Total

Underdiagnosis:

Cardiovascular Cerebrovascular

Respiratory Digestive

4

1 1 2

1 9

I II

Infective Neoplasia

VIIa

Cardiovascular

VIII

Respiratory

3

IX

Digestive

5

X

Genitourinary

5

Others Total

1 6

25

1 46

Stomach carcinoma (2), colon carcinoma (2) Ischaemic heart disease (1) Cerebral infarction (1),

Of the 68 cases in which all diagnoses were not concordant, there were 10 in which the correct diagnosis, if achieved, would have led to changes in treatment. In six of these 10 the correct diagnosis, if suspected, would have led to different investigations. In two of the 10 cases it was felt that the correct diagnosis could not have reasonably been achieved in life. Three examples of cases with clinically important discrepancies are presented in table 6.

Bronchopneumonia (1), pulmonary fibrosis (1) Peptic ulcer (1)

DIAGNOSTIC CONFIDENCE, AGE, AND DURATION OF HOSPITAL STAY

Necropsy contrbutory condition Psoas abscess (1) Lung (1), prostate (2), oesophagus (1), Hodgkin's disease (1), metastases (1) Hypertension (7), ischaemic heart disease (5), pulmonary embolus (4), aortic stenosis (2), others (7) Pulmonary fibrosis (1), emphysema (1), acute bronchitis (1) Gastrocolic fistula (2), gastric ulcer (1), others (2) Renal calculus (2), pyelonephritis (2), hydronephrosis (1) Plastic bag bezoar (1)

The correlation of diagnostic certainty recorded by the clinician with agreement of clinical and necropsy diagnoses is presented in table 7. There was a significantly higher proportion of correct diagnoses where clinical confidence was high. However, even in this category, 6% of main diagnoses, 22% of contributory conditions, and 12% of causes of death were incorrect. No clinically important differences in diagnostic accuracy were found in relation to patient age and duration of hospital stay.

577

Prospective study of necropsy audit of geriatric inpatient deaths Table 5 Over- and underdiagnosis of cause of death by ICD codes with common diagnoses Category

Number

Clinical cause of death

I VIIa

Infective Cardiovascular

2 11

VIII X Total Underdiagnosis: I II VIIa

Respiratory Genitourinary

15 1 29

Septicaemia (2) Pulmonary embolus (7), acute myocardial infarction (3) congestive cardiac failure (1) Bronchopneumonia (15) Chronic renal failure (1)

Infective Neoplasia Cardiovascular

1 2 12

Necropsy cause of death Septicaemia (1) Liver (1), cholangiocarcinoma (1) Pulmonary embolus (8), congestive cardiac failure (1), arteritis (1),

VIII

Respiratory

4

Bronchopneumonia (2), aspiration

ICD Code

Overdiagnosis:

acute myocardial infarction (1), left ventricular failure (1)

pneumonia (2)

Total

19

Table 6 Examples of cases of clinically important discrepancies Case 1: Main diagnosis

Contributory condition Cause of death Others Case 2: Main diagnosis

Contributory condition

Cause of death Others Case 3: Main diagnosis

Contributory

Clinical findings

Necropsy findings

Bronchopneumonia Cerebrovascular disease Bronchopneumonia

Aspiration pneumonia Cerebral infarction Plastic bag bezoar Aspiration pneumonia Cystitis Chronic myocardial ischaemia

Carcinomatosis Septicaemia Bronchopneumonia Rheumatoid arthritis Septicaemia Nil

Septic arthritis Paravertebral and psoas abscess Rheumatoid arthritis Nil

Pulmonary embolus

Acute gastric ulcer with

Nil

condition Cause of death

Pulmonary embolus

Others

Senile dementia

Septicaemia

haemorrhage Hypertension Bronchopneumonia

Acute gastrointestinal haemorrhage Senile dementia Meningioma

Table 7 Agreement of clinical diagnoses with necropsy diagnoses by diagnostic certainty, figures in parentheses are percentages in each diagnostic category

Diagnostic

confidence

Number

Certain Fairly certain Uncertain Total

32 48 20 100

appreciation of the importance of minor findings in elderly patients with multisystem disorders. In both these studies the spectrum of disease was similar to the current series, with ischaemic heart disease, congestive cardiac failure, pneumonia, cerebrovascular disease and neoplasia accounting for most of the deaths. Rates of diagnostic discrepancies between clinical and necropsy diagnoses for general series range from 4-39%. A previous review of necropsies from general medical and surgical wards in Edinburgh indicated a much higher rate of misdiagnosis in patients over 75 years of age, with 53% of main diagnoses not confirmed at necropsy. This contrasts with a significantly lower one of 28% for patients older than 75 from geriatric wards in the present series (2 = 10-624; p = < 0-01). In Manchester Puxty and his colleagues identified a similar discrepancy in patients over 85 years with the rates of misdiagnosis from the general medical and orthopaedic wards being twice those from the geriatric unit.'5 Though a variety of explanations are possible, it is tempting to suggest that very old people are more likely to receive a detailed assessment in a geriatric unit where medical and nursing facilities are more geared to their needs. The disease types giving rise to diagnostic difficulty were similar to those in the previous study, with ischaemic heart disease, congestive cardiac failure, pulmonary embolus, bronchopneumonia and neoplasia frequently being both over- and underdiagnosed. Two cardiovascular conditions were of particular interest. The first was undiagnosed hypertension as a contributory condition in seven cases; the second comprised two cases where senile calcified aortic stenosis was not diagnosed as the main diagnosis underlying left ventricular failure.

Necropsies are important in the accumulation of health care statistics, in undergraduate Disagreement on Disagreement and postgraduate training, in research and in on cause of Disagreement on contributory providing audit of clinical diagnosis and treatmain diagnosis (%) conditions (%) death (%) ment. Despite these important contributions 2 (6) 7 (22) 4 (12) hospital necropsy rates have declined in the 15 (31) 15 (31) 18 (38) 11 (55) 10 (50) 12 (60) United Kingdom since the 1950s, and a recent 28 32 34 study recorded a necropsy rate as low as 8.2%. 16 Several factors have been cited as contributing to this declining rate, including irrelevance -of necropsy findings to clinical management, delay in issuing reports, a Discussion decline in clinicopathological conferences and Few recent studies have specifically investigat- overconfidence in the precision of modem ed necropsy findings in the elderly.'2"-5 The investigative techniques.3 " current necropsy rate of 50% compares favourThis series has shown that the prime factor ably with one of 20% recorded in a previous in establishing higher necropsy rates is the level United Kingdom study by Puxty et al, of of clinical and pathological interest. Where necropsy findings in elderly patients from necropsies were requested on all deaths then a geriatric, general surgery, and orthopaedic satisfactory necropsy rate of 50% was achwards. 5 ieved. In the current series no satisfactory cause of The results of this study indicate a greater death was found in 1 % of necropsies. This accuracy of diagnosis in the elderly by specialcontrasts with the 4*5% found by Puxty et al ist geriatricians compared with non-specialist and the much higher rate of 26% found by units. The levels of misdiagnosis are, however, Kohn et al'4 in a series of 200 necropsies still high and are no cause for complacency. performed on patients older than 85. This During the course of the study much was discrepancy may be due to an increased learned from the vigorous audit of cases at the

578

Paterson, Dorovitch, Farquhar, Cameron, Currie, Smith, MacLennan

clinicopathological meetings. With the planned introduction of greater audit in clinical medicine'8 we feel that clinicopathological necropsy meetings have an important role to play in auditing the care of elderly patients. We thank the resident medical staff, pathologists, and technicians for their cooperation, Mrs Joyce Garson for secretarial work, and Mrs Sheila Paterson for help with data analysis.

1

2

3 4 5

6

Cameron HM, McGoogan E. A prospective study of 1152 hospital autopsies. I. Inaccuracies in death certification. Pathol 1981;133:273-83. Cameron HM, McGoogan E. A prospective study of 1152 hospital autopsies. II Analysis of inaccuracies in clinical diagnoses and their significance. Pathol 1981;133: 285-300. Harrison M, Hourihane DOB. Quality assurance programme for necropsies. Clin Pathol 1989;42:1190-3. Peacock SJ, Machin D, Duboulay CEH, Kirkham N. The autopsy: a useful tool or an old relic? Pathol 1988; 156:9-14. Mercer J, Talbot IC. Clinical diagnosis: a post mortem assessment of accuracy in the 1980's. Postgrad Med 1985;61:713-16. Reid WA, Harkin PJR, Jack AS. Continual audit of clinical

diagnostic accuracy by computer: A study of 592 autopsy cases. Y Pathol 1987;153:99-107. 7 Landefeld SC, Chren M, Myers A, Geller R, Robbins S, Goldman L. Diagnostic yield of the autopsy in a university hospital and a community hospital. N Engl J Med 1988;318:1249-54. 8 Goldman L, Sayson R, Robbins S, Cohn LH, Bettmann M, Weisberg M. The value of autopsy in three medical eras. N Engl J Med 1983;308:1000-5. 9 Bauer FW, Robbins SL. An autopsy study of cancer patients. JAMA 1972;221:1471-4. 10 Scottolini AG, Weinstein SR. The autopsy in clinical quality control. JAMA 1983;250:1192-4. 11 Cameron HM, McGoogan E, Watson H. Necropsy: A yardstick for clinical diagnoses. Br Med J 1980; 281:985-8. 12 HofmanWI. The pathologist and the geriatric autopsy. JAm Geriatr Soc 1975;23:1 1-3. 13 Ishii T, HosodaY, Maeda K. Cause of death in the extreme aged. Age Ageing 1980;9:81-9. 14 Kohn RR. Causes of death in very old people. JAMA 1982;247:2793-7. 15 Puxty JH, Horan MA, Fox RA. Necropsies in the elderly. Lancet 1983;i: 1262-4. 16 Anderson NH, Shanks JH, McCluggage GWG, Toner PG. Necropsies in clinical audit. J Clin Pathol 1989; 42:897-901. 17 McGoogan E, Cameron HM. Clinical attitudes to the autopsy. Scot Medy 1978;23:19-23. 18 Department of Health. Working for patients. (White Paper) London: HMSO, 1989.

Prospective study of necropsy audit of geriatric inpatient deaths.

To evaluate the accuracy of clinical diagnosis by specialists in geriatric medicine and to compare this with a previous study involving non-specialist...
626KB Sizes 0 Downloads 0 Views